Navigating ontology and AI in information management - and why we should probably just buy a sextant

Peter Drucker once said that management is the organ of society responsible for making economic resources productive. To me, whether the label is information management, records management, data management, knowledge management - whatever label we practice under - that is what we have to be about. We're about making the resource productive - making it produce more than it is now. We're expected to do this by bringing specialist expertise to play. Right now, the specialist expertise that seems to be fashionable is ontology and AI - those techniques/technologies are the ones that practitioners seem to be talking about the most. The interesting thing to me about these discussions is that I don't think they solve the problems we have.

We have lots of problems, but the core problem for me is that management aren't taught how to manage the relationship between the quality of the information they work with, and the outcomes that they can achieve - taught that they can manage it, and that actively managing it can produce better outcomes. It makes them really unsophisticated information users, and unsophisticated thinkers about information. It means that when they plan, organise, lead and control the work that needs to be done, they develop a system for performing the work that's relatively independent of the information - the information gets left to the individuals to decide what to do. The outcome is fairly predictable - everyone makes locally optimal decisions, and the organisation misses out on the opportunity to set priorities that reflect what the organisation wants. For me, the mess we're in with content sprawl is a direct result of the lack of sophistication in ordinary rank and file managers in organisations.

Like all good disciplines, when we have an economic resource and we're supposed to be the experts in it, we want to focus on it - so we're spending a lot of time talking AI and ontology at the moment. Lets be clear, these are things that I love - undisciplined ontology feels as natural as breathing to me, and disciplined ontology has huge potential, I think AI also has huge potential in helping us deal with the unglamorous side of the work we do and helping people make sense of the world faster than ever before. What I think we need to recognise though, is that they're not going to solve the problem we actually have in information management - because they're a tool, and to make effective use of them, the way work is organised around them needs to change. That's not something we can do - that's something that organisational managers need to do. To get the best out of these new information technologies, their planning, organising, leading and controlling needs to consider how the social system of getting the work done meets the technical systems - technical systems which now include ontology and AI.

The social system that gets work done in organisations now is not a sophisticated one when it comes to information. That's why we're in the huge mess we're in. When we add ontology and AI, we're taking an unsophisticated system that mostly hasn't been able to use a classification scheme or a taxonomy effectively, and we're adding more complicated technologies in the expectation that somehow our organisations will be able to operate at the more sophisticated level because they have more sophisticated technologies.

200 years ago, the sextant was the most sophisticated thing in navigation. It transformed shipping by allowing ships to travel direct routes across open ocean rather than needing to hug the coast and rely on landmarks they could see for navigation. The thing is, that if you took a ship that organised its work around hugging the coast - making short trips from small port to small port to get to and from destinations that were further apart - and put a sextant on it, now it was a ship that hugged the coast and carried a sextant. The whole operation actually got less efficient, because now the ship (which was probably perfectly fine) had an expensive piece of technology on it. To use the sextant effectively, you needed a navigator who was mathematically literate - trigonometry (which no sailor knew before) was now a part of the game. Now though, things have become even more expensive - you've taken a ship that didn't have a sextant, or anyone who was mathematically literate (expensive 200 years ago), and added a sextant and an ultra expensive crew member. Getting value out of the sextant meant practices had to change - the sextant allowed direct navigation, so voyages could now be organised around direct travel between ports with open ocean between them. Open ocean travel though, required different ship design to do it safely, and to make money out of it required ships to carry larger cargoes than before. It also required ships to carry many days, sometimes weeks or months of food and water to avoid having everyone starve to death or die of thirst. In short, it required a different system of management - the system that was good at carrying small cargoes over short distances in sight of land, where water and food could be had relatively quickly, and shelter could be found if weather was poor, taking that system and way of organising work, and adding a sextant to it just made the same work more expensive. It wasn't until the organisation and management of the work around the sextant changed, that the sextant became valuable.

Ontology and AI are the same.

In information, records, knowledge and data management, we love the technology and the technique. Adding these technologies and techniques to the same system of organising and managing work though, just produces more expensive systems of organising and managing work. To be effective, the organisation and management of the work needs to change. To change the organisation and management of work, we need ordinary business managers who are more sophisticated thinkers about information, and how work can be organised around it. This has historically been a real blind spot for all of our professions. People are messy, and organising and managing work is complicated on its own, certainty is hard to find, it's easy to see understand why we want to just throw some AI or ontology at our problems - it gives us something clear-cut, and easy to do. We also have a good blueprint here - IT have been just throwing the next technology at things for decades, something we spend a lot of time criticising them for, because they don't think about how the information needs to be organised for the technology to work effectively. It is somewhat ironic that we are now slipping into the same trap, and ultimately making the same mistake that lots of IT management do - not considering how the work needs to be organised around the technology and technique.

The title offered the sextant as a bit of a provocation, but a real one. If we aren't going to engage seriously with business management, and help them raise their level of sophistication in managing the relationship between the quality of outcome they achieve through their work, and the quality of information (records, data, knowledge etc.), we should probably just buy a sextant. A sextant is much cheaper than a new EDRMS, or an ontology or AI project, and it gives us something interesting and fun to talk about - which will probably be of greater benefit to morale than another technical or technology project that just adds to the bloat that's already there. The goal of all of our professions - whatever their label - is to take economic resources, and make them more productive. Just adding more technology and technique doesn't achieve that, the technology and technique can't be effective until the work is reorganised - and that's not something we can do, that's something we need to do with business managers. Until we learn that very hard lesson, we're going to continue to get the results we've been getting - regardless of the technology and technique we use.